基于深度置信网络的盘形滚刀磨损预测研究

Research on wear prediction of disc cutter based on deep belief network

为了准确预测全断面掘进机盘形滚刀的磨损情况,提高掘进机掘进效率,降低掘进成本,首先通过参考已有的滚刀磨损理论计算公式,得到了影响滚刀磨损的因素;然后以岩石单轴抗压强度、盘形滚刀安装半径、刀盘推进力、掘进机掘进速度和刀盘转速为输入节点,以滚刀的单位径向磨损量为输出节点,采用MATLAB软件,分别建立了基于BP神经网络和深度置信网络的盘形滚刀磨损预测模型;最后以广州地铁某区间盾构机掘进工程为实例,对预测模型进行验证,并与理论计算结果进行对比。结果表明:基于深度置信网络的预测模型具有更好的预测性能,能较为准确地预测滚刀径向磨损,可为实际工程中指导滚刀更换时机提供参考。

In order to accurately predict the wear condition of the disc cutter for the full-face tunnel boring machine,improve the tunneling efficiency and reduce the tunneling cost at the same time,the factors affecting the wear of the disc cutter were obtained by referring existing theoretical calculation formulas for cutter wear.Then with the uniaxial compressive strength of rocks,the installation radius of disc cutter,the thrust force of cutter head,the tunneling speed of the tunnel boring machine and the rotating speed of cutter head as input nodes,and the unit radial wear of the cutter as output nodes,MATLAB software was used to establish disc cutter wear prediction models based on BP neural network and deep belief network,respectively.Whereafter,taking a shield tunneling project in a certain section of Guangzhou Metro as example,the prediction model was validated and compared with the theoretical calculation results.The results showed that the prediction model based on depth belief network had better prediction performance and could accurately predict the radial wear of cutters,which could provide reference for disc cutter replacement in practical engineering.